|
|
|
Cover Story
|
| |
|
|
|
Making connections
CRM in
telecoms |
Campbell
McCracken reports.
| |
|
|
|
| |
After years of being a monopoly, telecoms is now
a commodity product. It is as easy to change supplier as
it is to change brands of soap powder. In fact it is so
easy that mobile users are no longer required to
register with their supplier - they can simply buy a
phone and start using it immediately.
This
is giving the telcos a headache as they fight to stop
the 'churn' in their customer base. One of the weapons
that they are using to stem the flow of defectors is
Customer Relationship Management (CRM) - using all the
information that they have in their various databases on
the customers to try to predict which customers are
likely to churn, and to create more attractive pricing
packages for them.
"The
main aim of telcos is to provide better services with
less costs," says Dror Pockard, president of Clarify,
the newly formed CRM division of information solutions
providers Amdocs. But before they can do that, the
telcos need the right tools. "You have to provide
assistance to enable the operators to retain as many
clients as possible." This includes tools that allow the
operator to score the probability for customer
churn.
|
LEFT:
Netonomy's director of worldwide product
marketing, Andy Holcombe CENTRE: John Thomson,
WhiteCross’s vice president of worldwide marketing
RIGHT: Dror Pockard, president of Amdocs'
newly formed Clarify CRM
division |
|
|
|
|
Case study Whitecross at ntl |
|
Ntl is an
amalgamation of franchises in the UK. Over the
years these companies have been bought out by one
company, resulting in different billing systems
and different information and management systems.
Usually all of these systems have different
standardisation of data, and it becomes very
difficult to get one true picture of the whole
group.
Ntl took the CDR
data for one month (about 300 million records)
from all the different billing systems and found
none of it matched up. "It was all in different
formats," says ntl's Hopewell. "Things were
missing from it, some did time duration, some did
the time in seconds and minutes, different people
put different leading numbers on the telephone
numbers - all sorts of things."
The second month was
more successful. Ntl wanted to interact more with
the data so they got Whitecross - a provider of
business and customer intelligence solutions
to put it on to its server, which allows very fast
processing of the information. "We could take all
the call records and say 'If we want to change the
pricing to offer a better pricing to an
international destination, what would the impact
to the business be?'" says Fraser
Hopewell.
Whitecross and ntl
also did a data mining exercise where they took
every call and every customer and plotted this
information against a number of parameters
including the hour the call was made, what period
(day, evening, weekend), where they called, the
duration and the cost. "We did it for eighteen
parameters," says Hopewell. "We put parameters
around each of the points in the database to
cluster them together. What we came out of it with
was this database with profiles of our customers
which is self-generated from the sorts of calls
they make.
"In a simple
example, we might get a group of customers who
make a high percentage of their calls to an
international destination. Once we know we have
callers with a high percentage of their calls to
an international destination we can promote our
international calling to them so that they take up
the call plan, get cheaper rates and generate more
revenue because they're phoning more."
The data mining
exercise generated about a dozen profiles and ntl
was able to design tariffs and promotions and
incentives to target those specific segments. This
saved the cost of doing a shotgun approach to
mailing and got a much higher response rate
because they were telling the customers about
something that they were actually interested in.
Ntl also ran
programs to measure the accuracy of their billing
systems. "We could look through and see where we
were undercharging our customers," says Hopewell.
"That was usually through a mistype in the billing
engine or it was an old tariff which hadn't been
updated. So what we were able to do was to clean
up our billing process a lot and make the billing
a lot more accurate and make sure we brought
people on to the right tariff."
"We did some churn
analysis - we looked at how customers were
spending on a monthly basis by call type (local,
national, international) and how that changed from
month to month." What ntl found was that there was
a large drop in the amount of revenue we got from
local calls. It contacted some of the customers to
find out why they were leaving ntl and found they
were leaving to use one of rival BT's talk
together packages which gives them free local
calls. "So we were very quickly able to reduce the
churn on the issue, and reduce the numbers of
those customers we saw leaving."
The positive effect
of using the analytical systems and services from
WhiteCross exceeded ntl's expectations. "Our
bottom line results have improved since we have
been working with WhiteCross," says ntl's
Hopewell. "Our calculations have shown that £40m
of that improvement is directly attributable to
the business-critical information provided to the
management of ntl by WhiteCross. There's lots of
areas where the money was saved, including
reducing churn, more effective targeting, by
ensuring that we charge the right rates to the
right customer, and ensuring that we have the
right customers on the right tariffs."
"The biggest one
going forward will be us developing tariffs
against cluster segments, against customer
clusters. That will probably be our biggest
revenue generator because we can offer a specific
tariff to a specific customer, which will make it
highly more attractive to take up the package."
"We're going to
design tariffs for ethnic minorities, based on
their calling profile so that when our sales
people go and knock on someone's door they can say
'Do you call (wherever)? If you do, I've got a tariff that exactly
suits you.' Just by doing that you suddenly get
your take-up rates increasing by 30% - 40%
straight away, because you're offering something
the customer actually wants rather than a generic
flat tariff where we've tried to capture
everything." |
Customer Detail
Records The
data that the tools need to work on is available in
abundance - the Call Detail Records (CDRs). "When
everybody makes a phone call they generate a CDR," says
ntl's head of telephony, Fraser Hopewell. "That CDR
tells us what number they were from, what number they
dialled, the duration of the call, when they made the
call, and the cost." Ntl enlisted the help of a provider
of business and customer intelligence solutions,
Whitecross, to process their records (see Case Study -
'Whitecross at ntl').
Amdocs’ databases for customer churn and
profitability not only uses CDRs but will use all
available information. "We are running a data warehouse
concept to bring data in from various systems - whatever
systems are available," says Amdocs' Pockard. "We don't
only need to use Amdocs systems or billing
systems."
Other pieces of information that could be used to
predict churn include the customer's address and his
employer - if there are employees working for the same
corporation who have churned already then there is a
higher probability that the customer will churn. "Even
calls to another operation's call centre," says Pockard.
"If you see such a call happen, it's a red
flag."
Churn
Prediction Once
telcos forecast that a customer, or cluster of
customers, is likely to churn, they will want to do
something about it. This means moving the customers on
to a pricing plan that is both attractive to them and
profitable to the telco. "We run 'what if' analyses
against the database and a number of new pricing
schemes," says Whitecross's John Thomson. "We're looking
to meet that double condition of a point where we can
save money for the customers and increase the margin to
the telco." This can be achieved by discounting calls
that are made at times, and to locations, that suit the
telco and help the telco get maximum throughput on its
infrastructure.
"It's all based around the question: what are the
infrastructure costs at this time? Is it cheaper for
people to be making calls to Brighton, or do [the
telcos] have international deals with other
international carriers where it's cheaper for them to
provide the service to Australia in the middle of the
night?" explains Thomson.
Amdocs has solutions to suggest pricing plans for
customers too. Its Campaign Manager software uses the
same database as the churn analysis software to propose
plans that will fit the client. Telcos can use this to
produce as many new plans as they need. "The biggest
number of call plans I've seen is 40,000," says Amdocs'
Pockard, "But to be rational it's somewhere between 500
to 2,500."
The Future - Customer
Managed Relations As
the mobile telecoms market matures, a complementary
process to CRM is emerging - Customer Managed Relations
(CMR). "What we let the customers do is manage their
accounts without the need to contact the call centre,"
says Netonomy's director of worldwide product marketing,
Andy Holcombe. "We believe that looking after yourself
entirely as a customer is far more effective than
putting a call centre in charge."
Netonomy has found that the more times you get
someone to interact then the more loyal they become.
That remains true even if the a problem is being dealt
with, provided it its resolved. "Every time you talk to
the service provider, providing they are helping you in
some way, you're locking yourself in for a greater
amount of time."
CMR
is not an alternative to CRM - the two processes are
complementary. CRM is about understanding the customer;
CMR is about empowering the customer and letting them
take charge.
In
the future, as the third generation of mobile services
(3G) becomes established, we'll see the two processes
working together more and more. Some 3G service
providers expect that customers will use CMR to manage
their relationships around 10 times per week, which is
almost 100 times more often than the average customer
today. That means more information will be available to
feed into CRM databases, which in turn means that the
service providers should be able to get a better
understanding of their customers and increase their
retention rates.
|